# Strange production vs. development behavior difference

**URL:** <https://www.ruby-forum.com/t/strange-production-vs-development-behavior-difference/183723>\
**Category:** Rails\
**Created:** [February 19, 2010, 4:51am UTC](https://www.ruby-forum.com/t/strange-production-vs-development-behavior-difference/183723 "2010-02-19T04:51:50Z")\
**Posts on this page:** 7\
**Page:** 1

<div class="post-metadata">

**Author:** ![pepe](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/p/4da419/32.png) [@pepe](https://www.ruby-forum.com/u/pepe)\
**Post date:** [February 19, 2010, 4:51am UTC](https://www.ruby-forum.com/t/strange-production-vs-development-behavior-difference/183723/1 "2010-02-19T04:51:50Z")

</div>

I am running into something that has me stunned.

I have a wizard like application that works perfectly well in  
development but does not in production. The code has nothing  
environment specific going on and the same DB is used in both modes so  
I know the problems are not being caused by the data.

The specific problem I have is that I am supposed to display a section  
of the page only if the same SSN is in the DB already. The SSN is  
already in the DB so the page should always display.

The funny thing is that the page works as expected for both creation  
and update in development but only for creation in production.

Furthermore, there is something really funky going on. If I add this  
at the beginning of my action:

```
render :text => 'in my action'
return

```

the development environment will execute it with no problem but the  
production environment completely ignores. It is like if I had made no  
changes to the code. And in case you were going to ask, yes I made  
sure (about 10 times) that the web server (Mongrel) had been restarted  
when running in production.

I have been looking around the whole day and I can’t find anything  
that would give me a clue about what the problem might be.

Any help would be greatly appreciated.

Rails is NOT installed in vendors  
Ruby: 1.8.6  
Gems:  
\*\*\* LOCAL GEMS \*\*\*

actionmailer (2.3.5, 2.2.2)  
actionpack (2.3.5, 2.2.2, 1.13.6, 1.13.3)  
actionwebservice (1.2.6, 1.2.3)  
activerecord (2.3.5, 2.2.2, 1.15.6, 1.15.3)  
activerecord-oracle-adapter (1.0.0.9250)  
activerecord-oracle\_enhanced-adapter (1.2.3)  
activeresource (2.3.5, 2.2.2)  
activesupport (2.3.5, 2.3.3, 2.2.2, 1.4.4, 1.4.2)  
archive-tar-minitar (0.5.2)  
builder (2.1.2)  
calendar\_date\_select (1.15, 1.13)  
cgi\_multipart\_eof\_fix (2.5.0)  
color (1.4.0)  
cucumber (0.6.2)  
diff-lcs (1.1.2)  
fastthread (1.0.1)  
fxri (0.3.6)  
fxruby (1.6.6)  
gem\_plugin (0.2.3)  
hoe (1.8.2)  
hpricot (0.4)  
ibm\_db (1.1.1)  
json\_pure (1.2.0)  
linecache (0.43)  
log4r (1.0.5)  
mongrel (1.1.5)  
mysql (2.7.3)  
paginator (1.1.1)  
pdf-writer (1.1.8)  
polyglot (0.2.9)  
rack (1.0.1)  
rails (2.3.5, 2.2.2)  
rake (0.8.7, 0.8.3)  
rspec (1.3.0)  
ruby-debug-base (0.10.3, 0.10.2)  
ruby-debug-ide (0.4.5, 0.3.1)  
rubyforge (1.0.3, 1.0.2, 1.0.1)  
rubygems-update (1.3.5, 1.3.1)  
sources (0.0.1)  
sqlite3-ruby (1.2.1)  
term-ansicolor (1.0.4)  
transaction-simple (1.4.0)  
treetop (1.4.3)  
win32-clipboard (0.4.1)  
win32-dir (0.3.1)  
win32-eventlog (0.4.3)  
win32-file (0.5.3)  
win32-file-stat (1.2.3)  
win32-process (0.5.1)  
win32-sapi (0.1.3)  
win32-sound (0.4.0)  
win32console (1.2.0)  
windows-pr (0.6.2)

---

<div class="post-metadata">

**Author:** ![pepe](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/p/4da419/32.png) [@pepe](https://www.ruby-forum.com/u/pepe)\
**Post date:** [February 19, 2010, 7:55am UTC](https://www.ruby-forum.com/t/strange-production-vs-development-behavior-difference/183723/2 "2010-02-19T07:55:35Z")

</div>

On Thu, Feb 18, 2010 at 7:50 PM, pepe [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote:

> production environment completely ignores. It is like if I had made no  
> changes to the code. And in case you were going to ask, yes I made  
> sure (about 10 times) that the web server (Mongrel) had been restarted  
> when running in production.
> 
> I have been looking around the whole day and I can’t find anything  
> that would give me a clue about what the problem might be.
> 
> Any help would be greatly appreciated.

Pepe, as of Rails 2.3.5, the difference between the production and  
development environments  
are as follows:

- caching (by default)

- reloading of models, views, and controllers on each request (by  
default)

Good luck,

-Conrad

---

<div class="post-metadata">

**Author:** ![pepe](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/p/4da419/32.png) [@pepe](https://www.ruby-forum.com/u/pepe)\
**Post date:** [February 19, 2010, 3:41pm UTC](https://www.ruby-forum.com/t/strange-production-vs-development-behavior-difference/183723/3 "2010-02-19T15:41:02Z")

</div>

Thanks Conrad,

I knew about the caching and reloading differences, that is the reason  
I have been stopping and starting the web server every time while  
debugging the production environment. My understanding is that the web  
server will work with the application as if it were the first time it  
“saw” it every time it is re-started. If that is true, why doesn’t it  
take the changes I make to my action? Is there something else I should  
do or look up that might be causing this behavior? Has anybody  
encountered a similar problem? If there is something kept “in memory”  
that might be causing the problem, where to look?

Thank you.

Pepe

---

<div class="post-metadata">

**Author:** ![pepe](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/p/4da419/32.png) [@pepe](https://www.ruby-forum.com/u/pepe)\
**Post date:** [February 19, 2010, 4:09pm UTC](https://www.ruby-forum.com/t/strange-production-vs-development-behavior-difference/183723/4 "2010-02-19T16:09:43Z")

</div>

These kind of errors normally belong to the “editing the wrong file”  
bag. Can you double-check that the render :text is there with less(1)  
in the production machine? That that tree is the one being served?  
That you are pointing your browser to the correct machine?

---

<div class="post-metadata">

**Author:** ![pepe](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/p/4da419/32.png) [@pepe](https://www.ruby-forum.com/u/pepe)\
**Post date:** [February 19, 2010, 5:33pm UTC](https://www.ruby-forum.com/t/strange-production-vs-development-behavior-difference/183723/5 "2010-02-19T17:33:26Z")

</div>

On Feb 19, 4:17 pm, pepe [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote:

> was probably being processed by Rails after the newer one was and the  
> definition of the class was most likely being overlaid.
> 
> What I don’t understand now is why when I changed the newer version of  
> the controller Rails took the changes in the development environment.  
> Shouldn’t the controller class have been always replaced by the older  
> version even in development? I wish I knew more about the internal of  
> Rails to understand this. There is so much to learn…

Classes are loaded in different ways in production versus development

- in particular in production it will load pretty much every file in  
your app on startup, whereas in development it loads things as needed  
(and this mechanism would never cause foo\_controller\_save.rb to be  
loaded (unless you wrote FooControllerSave in your app.).

Fred

---

<div class="post-metadata">

**Author:** ![pepe](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/p/4da419/32.png) [@pepe](https://www.ruby-forum.com/u/pepe)\
**Post date:** [February 19, 2010, 5:18pm UTC](https://www.ruby-forum.com/t/strange-production-vs-development-behavior-difference/183723/6 "2010-02-19T17:18:47Z")

</div>

Xavier,

I had thought about that long ago and I was sure I was working with  
the right application. I don’t have any other application with the  
same name anywhere in the system and the changes I was making to the  
views for debugging purposes were being processed just fine.

However… I had a copy of the old controller file in the controllers  
folder. I had given it the same name and appended ‘\_save’ at the end.  
And that is what the problem was. The old version of the controller  
was probably being processed by Rails after the newer one was and the  
definition of the class was most likely being overlaid.

What I don’t understand now is why when I changed the newer version of  
the controller Rails took the changes in the development environment.  
Shouldn’t the controller class have been always replaced by the older  
version even in development? I wish I knew more about the internal of  
Rails to understand this. There is so much to learn…

Your comment made me remember that the class name in both files was  
still the same and since classes are never closed in Ruby that should  
be causing the problem. I removed the older version of the controller  
form the controllers folder and now it all works as expected.

Thank you very much!

Pepe

---

<div class="post-metadata">

**Author:** ![pepe](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/p/4da419/32.png) [@pepe](https://www.ruby-forum.com/u/pepe)\
**Post date:** [February 19, 2010, 5:39pm UTC](https://www.ruby-forum.com/t/strange-production-vs-development-behavior-difference/183723/7 "2010-02-19T17:39:52Z")

</div>

Thanks Fred, that explains it.

Thank you very much to all of you for your help.

Pepe

On Feb 19, 11:28 am, Frederick C. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid)